investigators have postulated multiple
orders of coiling or folding of a fundamental nucleohistone molecule (1, 3).
Several models have been derived from
25 June 1973; revised 11 September 1973
s
low-angle x-ray diffraction studies, including: four DNA molecules packed
into a single nucleohistone fibril (4);,a
single DNA double helix and associated
Spheroid Chromatin Units (v Bodies)
proteins folded into an irregular superhelix 80 to 120 A in diameter and 45
Abstract. Linear arrays of spherical chromatin particles (v bodies) about 70 A in pitch (5); and a single DNAangstroms in diameter have been observed in preparations of isolated eukaryotic
protein fiber constrained into a supernuclei swollen in water, centrifuged onto carbon films, and positively or negatively helix 100 A in diameter and 120 A in
stained. These bodies have been found in isolated rat thymus, rat liver, and pitch (6). Ultrastructural studies have
chicken erythrocyte nuclei. Favorable views also reveal connecting strands about also yielded a profusion of models.
15 angstroms wide between adjacent particles.
Spreading of chromosomes on a Langmuir trough frequently yields fibrils
The packaging of DNA within common for metaphase chromosomes about 250 A in diameter, although difeukaryotic chromosomes continues to (1). The DNA concentrations within ferences due to tissue type, presence of
be a formidable structural problem. localized regions of interphase nuclei chelating agents, and method of dePacking ratios greater than 100/ 1 (DNA may approximate 200 mg/ml or more hydration and drying have been reported
length/chromatid length) are not un- (2). Acutely aware of this problem, (7). Direct adsorption of sheared
chromatin onto microscope grids has
revealed a network of fibers approximately 100 A wide with numerous side
branches 80 to 200 A in length (5).
Spraying of chromatin onto a grid yields
a network of fibers (8) and separated
filaments (20 to 30 A in diameter) containing numerous "nodular" elements
about 150 A in diameter (9). Thin
sections of nuclei and chromosomes reveal fragments of threads frequently
100 to 200 A wide (3, 10, 11). Bram
and Ris (5) regard the 250-A fiber as
a folding (or doubling) of a superhelix,
due to divalent metal ions, and interpret
the thin-section data as artifacts of
chelation by buffer ions. Lampert (12)
views the 250-A filament as a folding
of the superhelix of Pardon and Wilkins
(6), and explains the thin-section data
in terms of shrinkage due to fixation.
Despite this divergence of views, there
is a consensus that multiple levels of
coiling or folding are required to explain the observed variation in chromatin fiber widths.
We have attempted to visualize
chromatin structure by methods different from those cited above. Interphase
nuclei were isolated from fresh rat
thymus (13), rat liver (2), and chicken
erythrocytes (2), washed and centrifuged twice in CKM buffer (14) and
once in 0.2M KCI, suspended in 0.2M
Fig.
Chromatin fibers spilling out of ruptured nuclei. The degree of fiber swelling KCl at a concentration of approxiand the proximity of individual v bodies to each other varies within different regions mately 108 nuclei per milliliter, and
of a single nucleus. Scale bars, 0.2 ,um. (a) Rat thymus chromatin, positively stained diluted 200-fold into distilled H20.
with a mixture of 4 percent aqueous phosphotungstic acid and 95 percent ethanol Nuclei were allowed to swell for 10 to
(3: 7), rinsed in 95 percent ethanol, and dried in air. (b) Rat thymus chromatin,
1 percent in
negatively stained with 0.5 percent ammonium molybdate, adjusted to p.H 7.4 to 8.0 15 minutes, then made
formalin
6.8
to
7.0).
Fixation
(pH
with ammonium hydroxide. (c) Chicken erythrocyte chromatin, negatively stained as
in (b). Clustering of v bodies is most evident in (c), where groups of three or more proceeded for at least 30 minutes. All
are readily visualized. Connecting strands are most easily seen in (b).
operations, up to this point, were at 0°
(Cornell Univ. Press, New York, in press)].
Similarly, TYA may lack significant effects on
platelet aggregation in vivo.
18. Agents used to induce platelet aggregation
were adenosine diphosphate (Sigma), epinephrine (adrenaline chloride: Parke-Davis),
bovine thrombin (Parke-Davis), and saline
extract of human subcutaneous connective
tissue (8).
19. Supported in part by National Heart and Lung
Institute grant HE 14595. We thank J. Rogers
(Roosevelt Hospital) for technical assistance.
I.
330
SCIENCE, VOL. 183
to 4°C. Aliquots of the swollen and
fixed nuclei were centrifuged through
10 percent formalin (pH 6.8 to 7.0)
onto carbon-covered grids, rinsed in
dilute Kodak Photo-Flo, and dried in
air, a technique developed by Miller
arfd co-workers (15). When examined
after positive staining, chromatin fibers
could be readily visualized streaming
out of ruptured nuclei. Fibers were
often very long (about 8 ,um in length),
unbranched, and in parallel arrays, and
revealed irregularly distributed thick
and thin regions. Frequently, views of
chromatin fibers (Fig. la) show spherical particles, v bodies (16), 60 to 80
A in diameter, connected by thin filaments (about 15 A wide). Less stretched
regions of chromatin revealed apparent
packing of v bodies. Analysis of fiber
widths from positively stained preparations showed peaks at 75 to 100, 125
to 150, and possibly 225 to 250 A
(Fig. 2), consistent with the ranges of
fiber widths described earlier. Better
visualization of the v bodies and connecting strands has been obtained by
the use of negative stains (Fig. 1, b and
c). The thickened fiber regions were
seen to represent clusters of v bodies.
Measurements of diameters of v bodies
for the different tissues employed
yielded the following average diameters
and standard deviations: rat thymus,
83 + 23 A; rat liver, 60 + 16 A; and
chicken erythrocyte, 63 + 19 A. Connecting strands, for rat thymus chromatin, exhibited average widths of 15 ±
4 A. Figure 2 demonstrates that the
distribution of diameters of v bodies,
measured from negatively stained preparations, superimposes on the lowest peak
of fiber diameters calculated from positively stained materials.
For a number of reasons we believe
that this appearance of chromatin fibers
as "particles on a string" is related to
the native configuration and is not an
artifact of the preparative procedures.
Washing of isolated nuclei in CKM
buffer (14) and in 0.2M KCl appears
to remove some nonhistone but no
histone protein (17-19) although histone
migration along the DNA cannot be
eliminated. However, nuclei so treated
reveal the same spectrum of chromatin
fiber widths after fixation and thin sectioning (2) as those observed for fixed
and sectioned whole tissue (10). Swelling of nuclei in water leads to stretching and thinning of chromatin fibers,
as revealed after fixation in water and
thin sectioning (20), and the disappearance of several of the low-angle x-ray
25 JANUARY 1974
be
Bt
tn
L4)
U
4..
._
CD
0.
I.
I
I
Width (R)
Fig. 2. Histograms comparing widths of positively stained chromatin fibers (solid lines)
with diameters of v bodies (broken lines) from negatively stained preparations. Random sampling of fiber widths was obtained by superimposing a lattice of lines 1 inch
apart on the photographic print and measuring the width of any fiber intersecting the
grid lines. The positively stained preparations were rat thymus (N = 100 samples) and
chicken erythrocyte (N = 360). The diameters of v bodies were measured only when
their edges were clearly defined and not overlapping another v body. The preparations
for measurements of v bodies were rat thymus (N = 114) and chicken erythrocyte
(N = 200). A calibration grid (54,864 lines per inch) was photographed with each
set of micrographs, printed, and measured simultaneously with the sample photographs.
reflections (21). Since addition of divalent metal ions to water-swollen nuclei
and chromatin does produce essentially
normal low-angle x-ray reflections (2,
21) and a partial return of ultrastructural morphology (20), the structural
changes exhibit some reversibility. Fixation with glutaraldehyde does not markedly perturb low-angle x-ray reflections
(2), although similar data for formaldehyde are not available. In order to demonstrate that nuclear isolation and washing are not essential for visualization of
the chromatin units, fresh whole chicken
erythrocytes were disrupted in cold 0.3
percent Joy detergent, followed by
swelling, fixation, and centrifugation
onto carbon films, a method developed
by Miller and Bakken (22). Although
the spreading of chromatin fibers was
not as good as that shown in Fig. 1,
particles resembling v bodies were observed throughout the chromatin.
Assuming that the v bodies described
here represent a real packaging of nucleohistone (23), we calculate some
of their expected physical properties.
For an average particle diameter of 70
A, a spherical shape, and a partial specific volume of 0.68 cm3/g (24), we
estimate an approximate molecular
weight of 160,000 per v body. Further,
we assume that every v body has at
least one histone of each of the five
classes, and that the sum of their mo-
lecular weights is 84,000 (25). Therefore, the DNA would have a calculated
molecular weight of about 80,000 and
a total length of about 400 A, packed
into the spheroid particle 70 A in diameter. Thus, a packing ratio (DNA
length/particle diameter) of about 6/1
might be expected. If there is significant
dehydration and shrinkage of the v
bodies, the calculated particle molecular weight would have to be increased.
The dimensions of ribosomal particles
and subunits, measured by electron
microscopy, are roughly half of the
calculated hydrated volumes from hydrodynamic measurements (26). It
would be conceivable, therefore, for
each v body to contain two of each type
of histone molecule complexed with a
double-stranded DNA with a molecular
weight of about 160,000. Further packaging of the DNA might then represent
a folded or helical close packing of the
spherical v bodies under the influence
of metal cations and noncovalent interaction. Studies should be directed toward fragmentation of chromatin and
isolation of particles with properties
complementary to the v bodies.
ADA L. OLINs, DONALD E. OLINS
University of Tennessee-Oak Ridge
Graduate School of Biomedical
Sciences, and Biology Division,
Oak Ridge National Laboratory,
Oak Ridge, Tennessee 37830
331
References and Notes
1. E. J. DePraw, DNA and Chromosomes (Holt,
Rinehart & Winston, New York, 1970).
2. D. E. Olins and A. L. Olins, J. Cell Biol.
57, 715 (1972).
3. H. Ris and D. F. Kubai, Annu. Rev. Genet.
4, 263 (1970).
4. V. Luzzati and A. Nicolaieff, J. Mol. Biol.
1, 127 (1959); ibid. 7, 142 (1963).
5. S. Bram and H. Ris, ibid. 55, 325 (1971).
6. J. F. Pardon and M. H. F. Wilkins, ibid.
68, 115 (1972).
7. J. G. Gall, Chrotnosoma 20, 221 (1966); S.
L. Wolfe, J. Cell Biol. 37, 610 (1968); H. Ris,
in Handbook of Molecular Cytology, A. Limade-Faria, Ed. (North-Holland, Amsterdam,
1969), p. 221; B. R. Zirkin, J. Ultrastruct.
Res. 36, 237 (1971); A. J. Solari, Exp. Cell
Res. 67, 161 (1971).
8. R. F. ltzaki and A. J. Rowe, Biochim. Biophys. Acta 186, 158 (1969).
9. H. S. Slayter, T. Y. Shih, A. J. Adler, G.
D. Fasman, Biochemistry 11, 3044 (1972).
- 10. J. S. Kaye and R. McMaster-Kaye, J. Cell
Biol. 31, 159 (1969); H. G. Davies, J. Cell
Sci. 3, 129 (1968); A. C. Everid, J. V.
Small, H. G. Davies, ibid. 7, 35 (1970).
11. B. R. Zirkin and S.-K. Kim, Exp. Cell Res.
75, 490 (1972).
12. F. Lampert, Nature (Lond.) 234, 187 (1971).
13. V. G. Allfrey, V. C. Littau, A. E. Mirsky,
J. Cell Biol. 21, 213 (1964).
14. The buffer consisted of 0.05M sodium cacodylate, pH 7.5; 0.025M KCI; 0.005M MgCl2; and
0.25M sucrose.
15. 0. L. Miller, Jr., and B. R. Beatty, Science
164, 955 (1969); J. Cell Physiol. 74 (Suppi. 1),
255 (1969); 0. L. Miller, Jr., B. A. Hamkalo,
C. A. Thomas, Jr., Science 169, 392 (1970).
16. The Greek letter v is suggested to denote
these spheroid chromatin bodies. This notation was considered to be appropriate since
they are both new and nucleohistone. Alternatively, we suggest that they be referred to as
"deoxyribosomes" in order to emphasize
their particular particulate analogy with
ribosomes.
17. Solvents similar to these have been shown
to extract ribonucleoprotein particles and
nuclear sap proteins [H. Busch, Histones and
Other Nuclear Proteins (Academic Press, New
York, 1965)] including the RNA-containing
interchromatin granules (18, 19). Furthermore,
preliminary observations (D. E. Olins and
E. B. Wright, unpublished) indicate that proteins extractable from chicken erythrocyte
nuclei by washing in CKM buffer and 0.2M
KCI contain no detectable histones, as judged
by gel electrophoresis in buffers containing
sodium dodecylsulfate LK. Weber and M. J.
Osborn, J. Biol. Chenm. 244, 4406 (1969)].
18. A. Monneron and Y. Moul6, Exp. Cell Res.
51, 531 (1968).
19. A. Monneron and W. Bernhard, J. Ultrastruct. Res. 27, 266 (1969).
20. K. Brasch, V. L. Seligy, G. Setterfield, Exp.
Cell Res. 65, 61 (1971); A. L. Olins and D. E.
Olins, unpublished observations.
/21. R. A. Garrett, Biochim. Biophys. Acta 246,
553 (1971).
22. 0. L. Miller, Jr., and A. H. Bakken, Acta
Endocrinol. Suppl. 168 (1972), p. 155.
23. It is considered unlikely that the v bodies and
connecting strands described in this report
might correspond to nuclear ribonucleoprotein particles. lnterchromatin granules (200 to
250 A in diameter) and perichromatin granules (400 to 450 A in diameter) are considerably larger than the v bodies (- 70 A in
diameter) (19). Interchromatin granules would
be expected to be extracted during the washes
in CKM buffer (18). Furthermore, v bodies
are observed in considerably greater numbers
than would be expected for the known ribonucleoprotein particles.
24. G. Zubay and P. Doty, J. Mol. Biol. 1, 1
(1959).
25. J. H. Diggle and A. R. Peacocke, FEBS
(Fed. Eur. Biochem. Soc.) Lett. 18, 138
(1971). From their data we take the following molecular weights: F2C, 20,800; F2B,
14,400; F2A2, 16,800; F2A1, 19,000; and
F3, 13,000. The substitution of Fl histone for
F2C in the rat tissue does not appreciably
change the sum of molecular weights. Adding
one molecule of each histone yields a total
molecular weight of 84,000.
26. W. E. Hill, G. P. Rossetti, K. E. Van Holde,
J. Mol. Biol. 44, 263 (1969).
27. A preliminary report of this work was pre-
332
sented at the meeting of the American Society
of Cell Biology, Miami, 15 November 1973.
At the same meeting, C. L. F. Woodcock
reported the observation of similar spherical
particles in chromatin fibers. The authors thank
0. L. Miller, Jr., and B. A. Hamkalo for advice
and criticism and M. Hsie for excellent technical assistance. This work was sponsored, in
part, by the Atomic Energy Commission under
contract with Union Carbide
Corporation, and,
in part, by NIGMS grant 1 ROI GM 19334-01
to D.E.O. One of us (D.E.O.) is a recipient
of NIGMS research career development award
5 K04 GM 40441-03.
19 July 1973
s
Tolerance to Heterologous Erythrocytes
Abstract. Injection of a water-soluble nonantigenic fraction obtained from
lysed sheep red blood cells virtually abolishes the subsequent immune response
to the red cells. The suppression is systemic and appears to be serum mediated.
Our understanding of the mechanisms lack of response to SRBC has been
underlying the production of immuno- variously attributed to loss or inactivalogical unresponsiveness or tolerance tion of thymus cells, thymus-derived
has increased over the past few years, cells, bone marrow or bone marrowbut as yet there is not-and perhaps derived cells, or to combinations of these
there cannot be-a unified theory of cells (4). The possible functioning of
tolerance that would explain the data suppressor cell populations has also
obtained from embryonic or neonatal been proposed (4, 7).
transplantation tolerance experiments,
We have shown (6) that a superfrom tumor enhancement and blocking natant fraction (after centrifugation at
work, from experiments involving dis- 40,000g) obtained after hypotonic lysis
aggregated serum proteins, monomeric of SRBC contains material that can
flagellin, or pneumococcal polysac- reduce to about 10 percent of normal
charides; nor may we be able to present the subsequent response of adult mice
a clear rationale for simultaneously to SRBC. The tolerogenic material did
understanding both low- and high-zone not appear to be immunogenic as meatolerance, both long-lasting and short- sured by induction of direct or indirect
term tolerance, and both antigen- and plaque-forming cells or by production
antibody-mediated tolerance (1).
of serum agglutinins or hemolysins.
I olerance to sheep red blood cells
Reciprocal experiments with hemoly(SRBC) has been obtained both in zate preparations of sheep and horse
neonatal and adult animals by a pro- red blood cells indicated, moreover, that
longed injection schedule with massive the tolerogenic effect of this supernatant
doses of SRBC (2). More generally, is highly specific (6).
however, unresponsiveness has been inWe now describe experiments deduced by treatment with antigen in signed to characterize the nature of the
combination with cytotoxic agents, such tolerance to SRBC induced by our suas cyclophosphamide (3, 4). Solubilized pernatant fraction. The experiments
SRBC fractions have also been used show that this tolerance is not associin induction of tolerance (5, 6). There ated with a reduction or elimination of
is no agreement concerning the basis imr,iunocompetent cells, but rather that
for the induced tolerance to SRBC. it appears to be the result of a serumDepending on the schedule of injec- mediated bloc'king effect.
tions, the number of cells and the source
Sheep red blood cells from individual
of cells used in restitution experiments, sheep were purchased (ARS-Spragueand the timing of assessment of toler- Dawley, Madison); they were washed
ance or of abrogation of tolerance, the before use. Tolerogenic preparations
Table 1. Ability of 107 spleen cells from normal or tolerant donors to respond to sheep red
blood cells after injection into lethally irradiated normal or tolerant syngeneic host animals;
PFC, plaque-forming cells; S.E., standard error.
Host
Donor
Experiments
(No.)
Normal
Normal
PFC
+
S.E.
(No.)
per spleen
15
4573 ± 882
15
8878 ± 1666
13
2751 ± 796
3
Tolerant
Normal
Normal
Tolerant
Animals
3
18
402±
94
SCIENCE, VOL. 183